If you are still using MagicReply and good old chatbot scenarios, this guide is for you. We'll review what the benefits of migrating your setup to Hugo are, and how to easily and efficiently do so. Ready?

Hugo is a next-generation AI support agent, fully integrated with Crisp.

While conversational at its core, Hugo goes beyond answering questions. He understands customer intent and context to proactively guide users toward resolution — just like a human support agent.

Built to autonomously handle a large share of incoming conversations, Hugo can also perform real actions through your integrations, such as checking account details, tracking orders, or issuing refunds, always within the rules you define.

By engaging proactively and providing concrete guidance, Hugo helps resolve issues faster and frees up your team to focus on complex cases.

Haven't met Hugo yet? Check out official resources

Why migrate from MagicReply chatbot scenarios to Hugo

Hugo is the AI Agent fully integrated with Crisp. Simply put, AI Agents can do much more than traditional chatbots, do it better, and do it effortlessly.

Deploying an AI Agent should be quick and easy, and this is exactly what Hugo provides.

By being fully natively integrated with Crisp, you can set it up literally in minutes without investing unnecessary energy in managing and building scenarios from scratch.

On top of that, Hugo also brings many new benefits:

• Dedicated inbox for conversations handled by Hugo → Live free of any noise while Hugo handles your conversations
• Advances escalation system→ Hugo is able to route conversations back to your agents if human intervention is required
• Stronger models, deep integrations and full customization → You have full control over Hugo's behaviour, knowledge and scope of action
• MCP integrations for Hugo to use tools and perform actions → Hugo can lookup a user profile, retrieve an order, or perform any custom action
• Monitor Hugo's impact thanks to dedicated analytics → Hugo can evaluate the topic of conversations and provide insight on how they are being handled
• Dirt cheap to use → Hugo was designed specifically for one task: handling support queries at the best level. In average, Hugo costs less than $0.10 per conversation
• And much more → Hugo brings new layers of customizations, better knowledge training, the ability to start workflows and perform actions in different situations

While standard chatbot flows provide reliable ways to pre-qualify your users or give first response elements to your users, Hugo is here to actually resolve any customer request and take off a large chunk of your team's load, which is ultimately the goal for any business.

Hugo doesn't just save time, he proactively engages with your users and provide guidance to provide the best customer experience. He is there to ensure that users get an accurate response in a timely fashion and help them resume their journey on your platform more knowledgeable and confident about your products or services.

Ready to start using Hugo?

Starting the migration process

If you were still using Chatbot scenarios and MagicReply, there are a couple of steps required before getting started with Hugo. Don't worry though, this is quickly taken care of.

To get started with the migration, simply navigate to Crisp and click ⚡️ AI Agent on the left menu.

Once inside the AI Agent menu, you'll notice that all your scenarios are now already present by default in Automate → Workflows Builder, and still working just as you had formerly configured them.

You can also find that your existing training resources were also migrated under the Train section.

The next step of the setup can be devised in two catefories, depending on your use-case:

• A. AI-first → If your current scenarios mainly revolve around the usage of AI with MagicReply and deflecting conversations from your team
• B. Chatbot-first → If your current scenarios mainly aim at pre-qualifying users with questions, offering them choices, categorization or routing

Still unsure about which option fits you best? Shoot us a message, we can help figuring this out with you.

A. [Recommended] Migrating to Hugo with an AI-first approach

Implementing Hugo from scratch is the quickest and cleanest way to get started. If your setup already primarily revolves around the usage of AI, Hugo is able to almost entirely replace your scenario usage on his own, natively.

Hugo will automatically handle conversations and escalations for you. You essentially no longer need to use scenarios (now referred to as "workflows") for AI purposes.

✔ To complete the migration, simply follow our Getting started with Hugo AI guide.

If you had workflows dedicated to certain tasks (such as detecting certain types of requests, triggering a specific branch or workflow based on an intent...), you can simply create dedicated workflows starting with an Action block, and later create routing rules for Hugo to automatically trigger these workflows for you when he encounters these situations.

B. Migrating to Hugo with a Chatbot-first approach

Before proceeding, this is a good opportunity re-evaluate your current setup, in case it can be accommodated with the recommended option A instead.

Considerations

There are various reasons you might not want Hugo to automatically handle conversations, but in most scenarios, your use-case can likely be accommodated by Hugo and other Crisp features.

1. If you are using scenarios to pre-qualify or route conversations:

• If you using scenarios to assign or segment conversations, you can instead use the recommended Triage feature.
• When Hugo handles conversations, they are moved into a dedicated Automated inbox. If escalation is required, he will automatically route them back to your main inboxes, and trigger your usual Sub-inbox and Operator routing rules.

2. If you are using scenarios to tweak the delay after which AI can trigger or the channels MagicReply handles:

• You can configure the delay after which Hugo can trigger again directly from Hugo inside AI Agent → Agent → Settings
• You can also configure your office hours in that same menu, if you wish for Hugo to only trigger when your team is offline
• This menu also allows you to select the conversation channels Hugo should only handle

3. If you are using scenarios to collect information or offer choices:

This is a very popular use-case inherited from chatbot-like flows. While this is predictable and reliable, chatbots are also archaic and often a source of pain for your users.

It requires them to go through several steps, often tiresome and potentially irrelecant to their request, before being able to ask their question.

These flows generally aim at providing some response elements, orienting the user or pre-qualifying them before an agent takes over. But ultimately, your goal is to resolve the user's request and bring an answer to their question. This is what Hugo is designed to perform much more efficiently and faster than any precedural chatbots.

4. If your scenarios only bring minor optimizations for your support:

We would also recommend using Hugo first in that case. Keeping a chatbot-first flow for minor and inconsequential optimizations is often not worth the trouble.

Instead, you would be able to deploy Hugo as a first-responder to your conversations and not only gain in simplicity and maintainability, but also drastically improve the effiency and speed of the support received by your users.

Migrating with a chatbot-first implementation

Inside of your scenarios, you will simply need to adapt your current setup to use Hugo instead of MagicReply.

This will involve replace branches containing "MagicReply" by the dedicated and modern "Answer with Hugo" agent block. The "Answer with Hugo" block acts similarly to an exit block, it will end the current scenario and simply hand the conversation to Hugo, who will start handling it.

Hugo also handles the escalation process based on the rules you've configured in AI Agent → Agent → Settings. This means that you no longer need to implement a "fallback" branch, nor buttons/options after Hugo sent a response.

✔ We have prepared an example to help you understand how to use this "Answer with Hugo" block in this dedicated guide.

Frequently Asked Questions

Still have some questions which were not covered above? Perhaps you'll find our answer in this FAQ.

If your question does not figure here, share it with us directly by chat, we'll be available to assist you

Will the migration incur downtime?

None whatsoever. When migrating, your current setup is automatically imported and continue working exactly as it currently does.

Your existing scenarios and training data are automatically enabled into the new AI Agent section, and remain enabled just as they previously were. This allows you to implement Hugo features and adapt your setup at your own pace, without any risk of disrupting your pre-existing implementation.

Will I begin to be charged when migrating?

Nope, your setup will continue to run as is, even if you are using MagicReply.

Charges would only come into effect once you activate Hugo to handle conversations, either when:

• implementing "Answer with Hugo" agent blocks into your scenarios
• Activating the Enable Hugo AI Agent on incoming conversations option in AI Agent → Agent → Activation.

Note that all Crisp paid subscriptions include free AI usage quotas depending on your plan. You can find more about those on our pricing page.

What if I don't want to use Hugo?

Hugo provides many benefits, but the choice is yours, and you can still benefit from many features even without activating Hugo.

Once you've migrated your setup, you can take the time to test Hugo inside the Playground. You also get some free AI credits on all plans to try it out live!

In which cases should I avoid migrating?

Absolutely none.

Migrating will only provide advantages. Go at your own pace and opt to migrate when you are ready.

By migrating, you'll for instance unlock additional domain crawling, improved ways to manage your training resources, and additional options to customize your AI's identity and behaviour.

Can I test it first?

Of course! Once you've migrated, your setup will work as usual, and you'll also unlock a special playground area to test Hugo at will before adopting him.

The playground not only allows you to test Hugo before deploying it, you can also use it to test your legacy scenario-based setup

Updated on: 03/02/2026